The other day I watched the webcast of Light Reading on The Role of AI in Platforms for Future RAN Systems [1], with NVIDEA, Fujitsu and Supermicro sharing their thoughts and market vision. Below you will find my interpretation of this webcast.

The telecommunications industry is rapidly evolving, and the integration of Artificial Intelligence (AI) in Radio Access Networks (RAN) is at the forefront of this transformation. While the committees responsible for defining 6G standards have agreed that 6G will be AI-native, many operators are not waiting until 2030 to reap the benefits. Instead, they are already envisioning the addition of AI to their current RAN installations. By doing so, they can leverage AI’s advantages in optimizing network operations, enhancing performance, and improving operational efficiencies.

5G networks have seen the implementation of Machine Learning (ML) to optimize power usage, support predictive maintenance in base cells and central offices, reduce operating costs, and improve service levels. However, the major use case for RAN in 5G has primarily been driven by commoditization rather than monetization of the added intelligence. ML was initially expected to create new monetization opportunities for 5G, but the primary benefits have been increased bandwidth and reduced latency to support high-demand applications such as video on demand, live streaming, and immersive VR/AR experiences.

The synergy between AI and RAN is set to revolutionize the telecommunications landscape, paving the way for more intelligent and efficient networks. By integrating AI into current RAN installations, operators can unlock new opportunities for monetization and further improve the efficiency and performance of their networks.

AI Opportunities in RAN

Operational Improvements

AI can significantly enhance and automate RAN operations. With AI-defined networks, the coding and configuration of the network can be optimized dynamically. This means that AI can continuously analyze network performance and make real-time adjustments to improve efficiency. Currently, there are hundreds of parameters that need to be set for RAN base stations, many of which are still manually configured or set by ML due to their complexity. AI can take this a step further by automating the optimization of these parameters, leading to more efficient and reliable network performance.

AI-Defined Networks

AI-defined networks represent a paradigm shift in how networks are managed and optimized. By leveraging AI, networks can become self-optimizing and self-healing, reducing the need for manual intervention. AI algorithms can analyze vast amounts of data from the network and make intelligent decisions to optimize resource allocation, improve signal quality, and minimize latency. This not only enhances the overall performance of the RAN but also ensures a more consistent and reliable user experience.

Parameter Optimization

One of the significant challenges in RAN management is the configuration of numerous parameters for base stations. These parameters need to be fine-tuned to achieve optimal performance, which can be a complex and time-consuming process. AI can simplify this by automatically optimizing these parameters based on real-time data. This ensures that the network operates at peak efficiency, even as conditions change. By automating parameter optimization, AI can help reduce operational costs and improve service levels.

Monetization Strategies

As AI is deployed to the edge of the RAN network, telco operators need to provision for the worst-case scenarios to ensure network reliability. This often results in underutilized AI resources during periods of low demand. To address this, service providers can explore ways to monetize the AI capacity by making it available to customers. This could include offering High-performance AI (HAI) as a Service, GPU as a Service, and leveraging High Compute Nodes for AI inference tasks. Currently, 14 major telcos worldwide already offer AI as a Service, with T-Mobile US leveraging their central offices as shared AI/RAN facilities.

Benefits of AI at the Edge

Deploying AI at the edge of the RAN network offers several advantages. It brings AI capabilities closer to end users, reducing latency and improving response times. This is particularly beneficial for applications that require real-time data processing, such as autonomous vehicles, augmented reality (AR), and virtual reality (VR). Additionally, edge AI can enhance data privacy and security by processing sensitive information locally, rather than sending it to centralized data centers.

AI Requirements for Edge Computing

The increase in AI requirements at the edge is driven by AI-enabled devices, such as the recently introduced Apple Intelligence on the iPhone 15 Pro and iPhone 15 Pro Max. These devices require significant computational power for on-device AI processing, which necessitates robust edge computing infrastructure. As more AI-enabled devices enter the market, telco operators must ensure their base cell and central office (CO) infrastructure is ready to support the rollout of AI networks.

Conclusion

The integration of AI in RAN presents numerous opportunities for improving network performance and operational efficiencies. By automating network management and optimizing parameters, AI can help telco operators deliver better service levels and reduce costs. Moreover, finding ways to monetize AI capacity at the edge can unlock new revenue streams and ensure that AI resources are utilized effectively. As the industry continues to evolve, AI will play a crucial role in shaping the future of telecommunications. Is your base cell and CO infrastructure ready to support the rollout of AI networks?

Happy cabling and always happy to receive your opinions and views,

Dave

[1] The Role of AI in Platforms for Future RAN Systems (4821045)